Runbook #6 – Low Repair Percentage, Query and Audit Tracing

Cassandra provides audit logging and tracing features that allow administrators to capture and troubleshoot CQL requests’ execution. These features use server resources and can result in performance issues on the cluster, making it essential to disable them as soon as the required data is collected. To disable these features, administrators can follow different methods, including disabling audit logging in cassandra.yaml or using nodetool commands to disable tracing. Identifying whether these features are enabled can be done through nodetool commands, and administrators can diagnose slow-running queries by looking into Cassandra’s logs. Tracing can also be enabled on one or all nodes of the cluster to analyze query performance in the system_traces keyspace.

Excerpted from the text –“Audit logging is a feature in Cassandra that logs all successful or failed CQL requests.”